Journal of King Saud University – Science 32 (2020) 657–666

Contents lists available at ScienceDirect

Journal of King Saud University – Science

journal homepage: www.sciencedirect.com

Molecular docking and quantitative structure-activity relationship study

of anti-ulcer activity of quinazolinone derivatives
Muhammad baba Muh’d ⇑, Adamu Uzairu, G.A. Shallangwa, Sani Uba
Department of Chemistry, Faculty of Science, Ahmadu Bello University, P.M.B 1045, Zaria, Nigeria

a r t i c l e i n f o a b s t r a c t

Article history: Proton pump inhibitors portray the first choice for treating various ulcer diseases because it inhibits H+/
Received 9 April 2018 K+-ATPase enzyme, by covalently binding to a cysteine residue of either potassium or proton pump,
Accepted 8 October 2018 therefore this enzyme is a validated target for anti-ulcer drugs. A Quantitative structure-affinity relation-
Available online 11 October 2018
ship (QSAR) and molecular docking analysis were carryout on 30 quinazolinone derivatives as H+/K+-
ATPase inhibitors. QSAR study was performed using Material studio software version 8.0, while molecular
Keywords: docking analysis of all the novel quinazolinone derivatives was performed using Autodock vina version
K+/H+-ATPase
4.0 of Pyrx software. The QSAR result reveal a strong correlation value of R2 = 0.9131, R2adj = 0.8914,
GFA
PUD QSAR
Q2LOO = 0.8038 and R2 pred = 0.8946 which showed a highly predictive and statistically significant model.
DFT Molecular docking analysis revealed that the ligand 25 bind tightly deep to H+/K+-ATPase (protein target).
(B3LYP/6-31G⁄) Because of the high binding affinity of
9.3 kcal/mol. This research has revealed a significant correlation
between binding score and biological activities of the molecules, and the results are even better than the
one proposed by other authors, more also, both results of quantitative structure-affinity relationship
(QSAR) and docking studies agree with each other which give chance for design and synthesis of novel
anti-ulcer agents exhibiting good action against the receptor (H+/K+-ATPase) Furthermore, the present study
prove more potent drugs than the ones already marketed.
Ó 2018 The Authors. Production and hosting by Elsevier B.V. on behalf of King Saud University. This is an
open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction sue, with perceptible deep or involvement of the submucosa

(Konturek and Konturek, 2014). In another term, Peptic ulceration
An ulcer can be referred to a surface/region in the digestive sys- is due to an imbalance between the petulant factors, such as pepsin
tem where the tissue has been ruined or damaged by the gastric and gastric acid, and local mucosal defenses such as prostaglan-
juice or other digestive enzymes that are produced by the stomach dins, mucus, and bicarbonate. Other main causes of the disease
(Chaudhary et al., 2015; Shamsuddeen et al., 2009). The common are Helicobacter pylori infection, use of pain killer’s drugs, smoking,
major type of ulcer that mostly affects the global population is exposure to stress and alcoholic drinking (Wang et al., 2016;
the peptic ulcer diseases, which occurs in the stomach or small Konturek and Konturek, 2014). The most common signs of peptic
intestine, the secreted gastric acid by the stomach compartment ulcer diseases include abdominal pain, hematemesis (vomiting of
destroy the protective cover of the small intestine or the stomach blood), chest pain, episodic gnawing and loss of appetite. Pain is
(duodenum) giving rise to acid-related disorders (Utzeri and usually alleviated during the night hours when the pH of gastric
Usai, 2017). Peptic ulcer disease can also be regarded as a malfunc- juice in the stomach has increased (due to circadian changes),
tion that occurs in the protective cover of the gastrointestinal tis- the pain is locally attributed to empty stomach about 2---5hrs
(Wang et al., 2017; Anand and Wakode, 2017) (See Table 1).
Peptic ulcer disease is one of the life-threatening diseases that
⇑ Corresponding author.
affect a large population of the world, over the past two centuries
E-mail addresses: [email protected] (M.baba Muh’d), Saniuba10@yahoo.
with high morbidity as well death rate (Ya-Li et al., 2015; Rajesh
com (S. Uba).
et al., 2017). Approximately 500,000 people are affected by the dis-
Peer review under responsibility of King Saud University.
eases in the United States, new 4million cases of peptic ulcer com-
plications, 1.8% lifetime prevalence (8–14%) with annual costs of
$4.82 billion in the develop countries (Noor et al., 2017). Gastric
Production and hosting by Elsevier proton/potassium pump (H+/K+-ATPase) is a phosphoenzyme rest-

https://doi.org/10.1016/j.jksus.2018.10.003
1018-3647/Ó 2018 The Authors. Production and hosting by Elsevier B.V. on behalf of King Saud University.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
658 M.baba Muh’d et al. / Journal of King Saud University – Science 32 (2020) 657–666

Table 1 Table 1 (continued)

Shows the structures and the activity (PIC50) of the quinazolinone chemistry.
S/NO Structures pIC50
S/NO Structures pIC50
12b 4.49
1a 3.67

2a 3.70

13a 4.53

3b 3.81

14a 4.61
4a 3.87

5a 3.85 15b 4.65

6b 3.89
16a 4.85

7a 4.22
17a 4.82

8a 4.29

18b 4.99

9b 4.26

19a 3.76

10a 4.38

20a 3.79

11a 4.43

21b 3.74
 M.baba Muh’d et al. / Journal of King Saud University – Science 32 (2020) 657–666 659

Table 1 (continued) et al., 2013). But these drugs are associated with side effects such
S/NO Structures pIC50
as renal failure, relapses, kidney/liver dysfunction etc. More also
the antibiotic resistance of H. pylori. Blockage of prostaglandins
22a 3.77
by anti-inflammatory drugs are all serious problems. Therefore,
Demand in search and design of potent molecules having anti-
inflammatory, and anti-ulcer action with an improved profile is
still a necessity (Rakesh et al., 2016). The Schiff’s base family is
made up of natural products, possessing diverse biological func-
23a 3.83 tions, which include: development of agrochemicals, medicines,
fungicide bactericide, antivirals antioxidants, antiproliferative
and antimicrobial drug (Rakesh et al., 2016). In medicinal chem-
istry, quinazolinone derivatives are part of this family which play
a crucial role in reducing gastric toxicity, hence, based on the above
fact, quinazolinone derivatives would be used to determine it anti-
24b 3.88
ulcer action.
Quantitative structure-affinity relationship (QSAR) and molecu-
lar docking techniques are largely apply to produce active drugs,
and to provide knowledge on how the drug interacts with receptor
often both method accelerate drug discovery process (Agarwal
25a 3.88 et al., 2013), the widespread application of quantitative
structure-affinity relationship (QSAR) technique is also used to
confirm a relationship that exists between the compounds and
their determined biological experimental activities (Abdulfatai
et al., 2017). Molecular docking is employed to determine the bind-
ing compatibility of the active site residues (receptor) to specific
26a 3.84
groups (ligands) through the use of scoring function by estimation
of the probability of small molecule bind to a macromolecule
(Agarwal et al., 2013). In view of this findings, this study is under-
taken with the aim to search for effective and novel anti-ulcer inhi-
bitors, which can be achieved through the use of molecular
27b 3.82 modeling by using Genetic Function Algorithm (GFA) method and
to determine the experimental biological activity of the com-
pounds. More also the compounds (quinazolinone derivatives) will
be docked against the H+/K+-ATPase enzyme (protein target).

28a 3.85 2. Materials and method

2.1. QSAR studies

2.1.1. Dataset collection

A series of 30 quinazolinone derivatives and their anti-ulcer
29a 3.79 activity against H+/K+-ATPase enzyme, were collected from litera-
ture and used for this study (Rakesh et al., 2016). The anti-ulcer
activities of these molecules measured as IC50 (lM) were normal-
ized and expressed as logarithmic scale as pIC50 (pIC50 = log1/IC50),
pIC50 was chosen to be the relying variable, while the descriptors
were selected as independent variables. Using Genetic function
30a 3.84 algorithm. The 2D structures and the pIC50 (anti-ulcer activities)
of these compounds are shown in Table.

2.1.2. Geometry optimization

ChemDraw Ultra 12.0 software was used to draw the 2D struc-
The letter a = training set while b = test set. ture of the compounds and were save as cdx file format. The struc-
tures were then converted to 3D using Spartan 14.0 version 1.1.2
software, the calculation was carried out using molecular mechan-
ing and concentrated in the parietal cells, is responsible for the ics force field (MM n+) to minimize the energy of the molecules
excess secretion of gastric acid into the stomach lumen, leading prior to the quantum chemical calculations.
to acid-related disorders (Agarwal et al., 2013). Therefore, this 6-311G* basis set, of density functional theory (DFT) using the
enzyme is unique to the parietal cells, it is considered as a good B3LYP method, was employed for whole geometry optimization
validated hit for anti-ulcer agents (Agarwal et al., 2013) because of the structures to obtain the lowest energy for all the compounds
the proton pump inhibitors reduce acid secreted by the stomach in this study (Abdulfatai et al., 2017). The DFT method uses Becke’s
via restraining the function of the enzyme. three-parameter functional (B3) with gradient correlation func-
There are currently many anti-ulcer drugs in the markets, such tional of Lee, Yang and Parr (LYP) which integrates a mixture of
as raperazole, lansoprazole, and omeprazole (Drini, 2017; Agarwal HF with DFT exchange terms.
660 M.baba Muh’d et al. / Journal of King Saud University – Science 32 (2020) 657–666

2.1.3. Molecular descriptors calculation Table 2

The 1D, 2D and 3D descriptors from the optimized structures of General minimum recommended value for the evaluation of QSAR model.

the Spartan files saved as sdf file format were generated using Symbol Name Value
Padel descriptor software version 2.18, this is because the Padel R2
Co-efficient of determination 0.6
descriptor software recognizes only the sdf file format (Yap, 2011). P(95%) Confidence interval at 95% confidence level <0.05
Q2 Cross Validation Co-efficient 0.5
2.1.4. Dataset division R2- Q2 Difference between R2 and Q2 0.3
N(ext&testset) Lesser number of external and test set  0:5
The biological data set was divided into a training and test set, R2ext. Co-efficient for determination of external and test set 0.6
in such a way that 70% (21) of the data set was made up of the
training set, while 30% (9) of the biological data was used as the
test sets. Kennard-Stone Algorithm was used for the division of
the biological data into a training set and test set (Kennard and where Yobs, Ypred, and Ymntraining are the empirical/experimental
Stone, 1969). activity, the second is the predicted activity while the latter repre-
sents the mean experimental activity of the sample in the training
2.1.5. Model building set, respectively (Abdulfatai et al., 2017).
The model was build using material studio software at Genetic R2 cannot be a useful measure for the goodness of model fitness.
Function Algorithms (GFA), the generated descriptors from Padel Therefore R2 is adjusted for the number of explanatory variables in
software tool were submitted for regression analysis with the the model, adjusted R2 (R2 adj) value varies directly with the
pIC50 values selected as the dependent variable while the descrip- increase in the number of repressors i.e. descriptors. Thus, the for-
tors were chosen as independent variables. The regression equa- mula for adjusted R2 is explained below:
tion was 4 which represent the number of the descriptors, 1000
ð n
1Þ ðn
1ÞðR2
PÞ
was chosen for the Population, and 500 was input for the Genera- R2 ¼ 1
ð1
R2Þ ¼ ðdÞ
tion parameter. The mutation probability was 0.1, and the number n
p
1 n
pþ1
of top equation returned was 4, using 0.5 as the smoothing param- The letter n is the number of training set compounds, p is the
eter. Friedman’s Lack of Fit (LOF) was used to score the model and number of independent variables in the model (Brand and Orr
other statistical parameters such as correlation coefficient matrix 2015).
(R2) for the internal, and R2prd for external validation, statistical sig- The (Q2) known as leave one out cross validation coefficient is
nificance was determined using F test (Fischer’s value); Q2 (cross- written as follows:
validated correlation coefficient). Below is the Friedman’s lack of P 2
fit formula. Y prd
Y
R2 ¼ 1
P  2 ðeÞ
C þ dp Y
Y mntrng
LOF ¼ SSE=ð1
ðaÞ
M where Yp represent the predicted activity, and Y represent observed
C refers to the number of terms in the model, other than con- activity of the training set, and Ymntrng is the mean activity value
stant time, while SSE is the sum of square of errors, p is the total of the training set (Abdulfatai et al., 2017; Jalali-Heravi and Kyani,
number of descriptors, while d is a user-defined smoothing param- 2004).
eter, p is the total number of descriptors contained in all model
terms, and M is the number of samples in the training set. 2.1.8. Y-randomization test
The structure of the regression model looks like this (David MLR models are built by randomly moving the activity while
et al., 2018): keeping the descriptors unchanged in Y-randomization test. The
R2 and Q2 values for the new QSAR models built for many trials
Y ¼ a1 x1 þ a2 x2 þ a3 x3 þ b ðbÞ
are expected to be very low, which assure that the developed
Y is the activity (pIC50), where ‘a’s and ‘x’s are regression coef- quantitative structure-affinity relationship (QSAR) models are
ficients for a conforming Nonpartisan variables representing strong. The c??2?? parameter calculated should also be greater than
molecular descriptors of the molecules, the ‘a’s correspond to ‘x’s. 0.5 so that the model will pass the test before recommendation.
While the last variable ‘c’ is the regression constant.  1=2
C
RP ¼ R  R2

ðaverage Rr Þ2
2.1.6. Quality assurance of the model
The quantitative structure-affinity relationship (QSAR) model
exploits in this study, was to evaluate the reliability and fatal capa- 2.1.9. Applicability domain
bility of the model, through the use of internal and external valida- A quantitative structure-affinity relationship (QSAR) model is
tions parameters. an essential statistical tool used to determine whether a model
make a good prediction within its applicability domain, and this
2.1.7. Internal and external validations can be determined from Williams plot (Tropsha et al., 2003). There
The standard used to compare internal and external validation are some techniques for assessing the suitable space of a QSAR
parameters of a particular quantitative structure-affinity relation- model, leverage is one of them and is given for a chemical com-
ship QSAR model is shown in Table 2 (Abdulfatai et al., 2017). Com- pound as hi:
ponent of the total variation assign to the model is known through  
1
the square of the correlation coefficient (R2) parameter. R2 is com- hi ¼ xi X T X xTi ði ¼ 1; . . . ; mÞ ðfÞ
monly used for internal validations, and the closer the R2 to 0.1 the
The xi represents the row-vector of the compounds’ X is a num-
better the regression equation tell us more about Y variable. The
ber of times constant descriptors matrix of the training set com-
expression for R2 is given below:
pound. It’s used as the prediction tool of the warning leverages
P 2
Y obs
Y prd (h*) emulating the limit for X values. Thus (h*) is given below:
R2 ¼ 1
P  2 ðcÞ
Y obs
Y mntrng pþ1
h ¼ 3ð Þ ðgÞ
n
 M.baba Muh’d et al. / Journal of King Saud University – Science 32 (2020) 657–666 661

The small letter is the number of training set compounds, while was found to be 0.8949 which is better than the standard c??2??
p is the number of descriptors that will be used to generate the parameter value 0.50 for the model.
model. Williams plot is the plot of standardized residual against From the above plot of the developed model, we can say that
the leverages, which give information on the pertinent surface of the model is stable and robust, and this is because of the high lin-
the model in terms of chemical range. The compounds that lie out- earity of the plot (straight line graph). More also the errors propa-
side the chemical domain (standardized residual no greater than 3 gated on both side of the zero indicate the strength of the model,
standard deviation units) are known as Y influential, while those the prediction of the test set data were determined using the
compound that are above the chemical domain are called outliers. developed model equation. And are shown in Table 3. The plot of
predicted values comprising both training set and test sets com-
2.2. Molecular docking studies pounds against the experimental pIC50 values are shown in
Fig. 1, the studied predicted activity is in accordance with the
Protein-Ligand docking studies of some quinazolinone deriva- experimental activity. Fig. 2 illustrates the plot of standardized
tives was evaluated in other to investigate the interaction between residual values versus leverages values (pIC50) (See Table 4).
the active site of H+/K+-ATPase enzyme and the ligands on Hp G630 Based on the result obtained in Table 5 above the novel quanti-
computer system, with Intel Ò CoreTM i3 Dual CPU, M330 @2.13 GHz tative structure-affinity relationship (QSAR) models generated pos-
2.13 GHz, 4 GB of RAM using Auto dock vina 4.2 of pyrex virtual sess significantly low predicted R2 and Q2 (leave one out) values for
screening software, Chimera version 1.10.2 and Discovery studio several tests and also CR2P value is higher than 0.5 which means the
software. Saccharomyces cerevisiae isomaltase crystallographic model is a powerful one and it is not obtained by probability.
structure (Resolution 1.30 Å PDB code 3AJ7,) with 72.4% succession
identity with the target was used as the mold for this study.

2.2.1. Ligands preparation for docking.

The 2D structure of the compounds (quinazolinone derivatives) Table 3
was drawn using ChemDraw Ultra 12.0 software, and where later List of the descriptors, their description, and classes for model 1.

converted to 3D structures for geometry optimization of the com- S/ Name Description Class
pounds, using Spartan’14 version 1.1.2, PaDEL Descriptor version No
2.18 (Yap, 2011; Anonymous, 2013). 1 BCUTw-1h Nlow highest atom weighted BCUTS 2D
2 BCUTp-1h Nlow highest polarizability weighted BCUTS 2D
3 MLFER_BO Overall or summation solute hydrogen bond 2D
2.2.2. Preparation of receptor basicity
The structure of gastric proton pump inhibitors, with the PDB 4 WTPT-4 Sum of part lengths starting from oxygen 2D
code 2Zex receptor, was downloaded from Protein Databank
(PDB). The 3D structure receptor was prepared by discarding water
molecules and cofactors using Discovery studio software (Ravin- 6
chandran et al., 2011) and save as Pdb.
5
Predicted acvity

2.2.3. Docking of the ligands with the receptor using Autodock version 4
4.0 of pyrx software.
3
The docking of the ligands (quinazolinone derivatives) and the
receptor (H+/K+-ATpase) was perform using Autodock version 4.0 2
of pyrx software (Trott and Olson, 2010). Chimera 1.10.2 software
1
was used to form the complex (ligand-receptor) since the receptor
and the ligand separate after carrying out the docking with auto- 0
dock vina of pyrx. The complexes were visualized to view their 0 1 2 3 4 5 6
interactions using Discovery studio software. Eperimental acvity
tranig set test set

3. Results and discussion Fig. 1. Plot of Predicted activity of both training set and test set versus
experimental activity for model 1.
3.1. Results of the QSAR.

3
Four quantitative structure-affinity relationship (QSAR) models
were developed using material studio software (Genetic Function
2
Algorithm). And out of these four models, model 1 was selected
Standardised residuals

as the preferable model for predicting the pIC50 for anti-ulcer com- 1
pounds, and this was based on the internal validation statistical
parameters of the model as it is in line with the recommended val- 0
idation parameters for QSAR models. The result of the model 1 is
-1
given below:
pIC50 = 0.021537762 * BCUTw-1h - 0.436958289 * BCUTp-1h
-2
+ 1.343771099 *MLFER_BO + 0.094105559 * WTPT-4 + 3.182054
R2trng = 0.913172, R2adj = 0.891465, Q2LOO = 0.803874, N trng = 21 -3
R2test = 0.894996, N test = 9 0 0.2 0.4 0.6 0.8 1 1.2
The high leave one out cross validation coefficient calculated Q2 Leverages
Trng Set Test set
LOO value (0.8038) for pIC50 shows a valid internal validation of
the model, the test set containing 30% of the biological data set Fig. 2. A plot of the standardized residual against leverages of both the training and
were used for the external validation of the model and the result test sets of model 1.
662 M.baba Muh’d et al. / Journal of King Saud University – Science 32 (2020) 657–666

Table 4
External validation of model 1.

S/No. pIC50 BCUTw-1h BCUTp-1h MLFER_BO WTPT-4 Yprd YPrd
Yobs

3b 3.81815 34.96938 9.353764 2.806 5.108277 4.099351 0.281194
6b 3.89619 15.99794 9.349122 2.612 9.937503 3.886542 -0.00965
9b 4.26760 15.99831 9.353553 3.032 7.889273 4.256248 -0.01136
12b 4.49485 15.99898 9.349303 3.016 10.41791 4.474578 -0.02027
15b 4.65757 15.99976 9.382758 3.041 13.51603 4.785121 0.127544
18b 4.99124 16.00001 9.349263 3.052 12.63542 4.731673 -0.26833
21b 3.74472 19.00061 9.353172 2.761 5.10812 3.695193 -0.04953
24b 3.88272 34.97016 9.914876 2.761 5.103419 3.793258 -0.08947
27 3.82102 78.91943 12.14228 2.819 5.108101 3.844922 0.023899

Table 5
Shows the result of the predictive R2 of model 1.

S/No. (Yprd-Yobs)2 Ymntrng Ymntrng Yobs- (Ymntrng
Yobs)2

3b 0.07907 4.0819 0.263744 0.069561
6b 9.32E-05 4.0819 0.185704 0.034486
9b 0.000129 4.0819
0.18571 0.034487
12b 0.000411 4.0819
0.41295 0.170528
15b 0.016267 4.0819
0.57568 0.331404
18b 0.072 4.0819
0.9181 0.842908
21b 0.002454 4.0819 0.337173 0.113685
24b 0.008005 4.0819 0.199171 0.039669
27b 0.000571 4.0819 0.260877 0.068057
R(Yprd-Yobs)^2 = 0.179000 R(Yobs-Ymntrng)^2 = 1.7047

Therefore R^2 = (1
(0.1790/1.7047) = 0.89499.

Table 6
Y randomization test.

Model R R2 Q2
Original 0.908394 0.825179 0.744396
Random 1 0.248977 0.061989 -0.44477
Random 2 0.407321 0.16591 -8.97625
Random 3 0.237331 0.056326 -0.93591
Random 4 0.391424 0.153213 -0.73844
Random 5 0.465703 0.21688 -0.1779
Random 6 0.544447 0.296423 -0.09683
Random 7 0.349944 0.122461 -0.44301
Random 8 0.2572 0.066152 -2.27191
Random 9 0.591068 0.349361 0.017999
Random10 0.184988 0.03422 -3.23752
Random Models Parameters
Average r: 0.36784
Averager2: 0.152294
AverageQ2:
1.73045
C 2
RP : 0.754499

Table 7 training and test set compound to fall within the applicability
Pearson’s correlation matrix of the descriptors in model 1. domain. Hence the model shows a very good prediction.
BCUTw-1h BCUTp-1h MLFER_BO WTPT-4
BCUTw-1h 1 3.1.2. Correlation matrix of the chemical descriptors from the best-
BCUTp-1h 0.953733 1 chosen model
MLFER_BO
0.28415
0.1893 1
A Correlation matrix was performed on the descriptors of model
WTPT-4
0.58115
0.4189 0.581646 1
1, and found to be highly correlated which means that the descrip-
tors used to build the model are very good. The result of the corre-
lation matrix is shown in Table 6 above (See Tables 7 and 8).
3.1.1. Williams plot for the best model
Fig. 2, above is known as Williams plot. It can be deduced from 3.2. Results of molecular docking studies of quinazolinone derivatives
the plot that four compounds were found outside the applicability
domain (h*= 0.5) for the model, three compounds with pIC50 of Molecular docking studies of 30 quinazolinone derivatives were
3.81, 4.65, 3.89 were the training set, while compound with pIC50 carried out and the docking scores of these compounds fall within
3.67 is the test set, indicating that these compounds are struc- the range of
7.2 to
9.3 kcal/mol. All the Compound were found
turally different from other ones and possess few chemical descrip- to strongly inhibit the H+/K+-ATPase enzyme by totally inundating
tors. Furthermore, the plot indicates that about 70% of both the efficient site in target protein, the result of docking analysis
 M.baba Muh’d et al. / Journal of King Saud University – Science 32 (2020) 657–666 663

Table 8
Binding energy, hydrophobic interactions, Electrostatic/other interactions, Hydrogen bonds and Hydrogen bond distance of some H+/K+-ATPase and the ligand.

Ligands Binding Energy Hydrophobic Interactions Electrostatic/Other Hydrogen Bonds Hydrogen Bond
(kcal/mol) Interactions Distance (Å)
7?
9.0 CLR3001, ILE36 ASP352 TYR158, GLU411, GLU277, and 4.27784, 5.49472
LEU791, TYR40 CLR3001 ARG442 3.86964, and
5.27925
19?
9.0 ASN284, ASP9 ARG72, ARG294, and ASP71, PHE12 GLN910, and PRO75 2.02787, 2.3628
PRO75, ILE913, and ALA74 ARG294 2.70481, 3.43443
3.48504, 3.0611
4.33, 3.5194
3.7856, 5.3713
5.37746 and 4.5958
21?
9.1 TRP988, TYR970 PHE959 3.44564, 4.09138
TRP988, TRP931 5.19086, 5.40558
TRP931, TRP988 TRP988, VAL26 4.12467, 4.64431
And VAL26 3.68473, 4.46669
5.12251 and 4.81825
23?
9.1 LEU800, TYR131 ASP128, THR804 ASP128, ARG979 2.86576, 2.41345
VAL805, ARG979 and ARG979 2.91186, 2.49163
ILE322, LEU800 3.4651, 3.59175
And TYR131 5.57346, 4.62663
3.9855, 5.43005
5.33404 and
4.29436
25?
9.3 TRP931, TRP931 3.8178, 3.73339
TRP988, TRP931 TRP931, TRP988 LEU958, 3.89195, 5.02769
ALA966 and VAL26 3.98529, 3.73676
4.82498, 5.0897
And 4.60284
26?
9.1 TRP931, TRP931 TRP988, TRP988 4.47939, 4.45599
TRP988, TRP931 TRP931, TRP988 LEU958, 3.7481, 3.94136
ALA966 and VAL26 3.98026, 4.67064
4.08594, 3.75553
4.79938, 5.06107
and 4.5588

Fig. 3. 2D and 3D structure of Ligand-Receptor complex 21 (
9.1 kcal/mol).

showed that all the docked ligands have lower energy value (high 21, 23 and 26, all with docked score of
9.1 kcal/mol, while 7,
binding energy value) compared to the standard anti-ulcer drug and 19 with docked score of
9.0 kcal/mol respectively, the docked
(omeprazole) with it binding energy value of
6.37 kcal/mol. More ligand 25 configuration display hydrophobic interaction with
also Figs. 3–6 depict the best low binding energy (high binding TRP931, TRP988, LEU958, ALA966 and VAL26 and two electrostatic
energy values) for the docked ligands. Among the 30 ligands that interaction TRP988, TRP988. These hydrophobic interactions indi-
were docked with the enzyme (H+/K+-ATPase), ligand 25 is the cated that ligand 25 bind deep in the core of active site where
most potent with the high docked score of
9.3 kcal/mol, ligands the reference ligand binds.
664 M.baba Muh’d et al. / Journal of King Saud University – Science 32 (2020) 657–666

Fig. 4. 2D and 3D structure of Ligand-Receptor complex 23 (
9.1 kcal/mol).

Fig. 5. 2D and 3D structure of Ligand-Receptor complex 25 (
9.3 kcal/mol).

 M.baba Muh’d et al. / Journal of King Saud University – Science 32 (2020) 657–666 665

Fig. 6. 2D and 3D structure of Ligand-Receptor complex 26 (
9.1 kcal/mol).

Furthermore, the high binding energy score of ligand 25 with correspond with each other and give direction for design of ulcer
other co-ligands was found to be better than those proposed by inhibitors. This study will help in design and development of a
the work of other researchers. drug which gives room for the synthesis of a new selective
H+/K+-ATPase inhibitor with predetermined affinity and activity
of the compound.
4. Conclusion
Acknowledgments
Quantitative structure-affinity relationship (QSAR) and molecu-
lar docking studies were carried out on 30 compounds of quina- The authors sincerely acknowledge department of chemistry,
zolinone derivatives as anti-ulcer compounds. The model from Ahmadu Bello University, Zaria for its technical support and Dr.
various physicochemical parameters corresponds to the essential Sani Uba and Mr. Usman Abdulfatai for their advice in the course
feature of the Schiff-base family (quinazolinone). The best model of this research.
with high correlation significant value (R2) of 0.8946 with active
site enzyme (H+/K+-ATPase), the molecular docking analysis of
References
quinazolinone analogs showed good binding affinity score of

9.3 kcal/mol against an enzyme responsible for the ulcer. The Abdulfatai, U., Uzairu, A., Uba, S., 2017. Molecular docking and quantitative
ligand was docked deeply within the binding pocket region form- structure-activity relationship study of anticonvulsant activity of
ing double-double hydrophobic interaction with TRP931, TRP931 aminobenzothiazole derivatives. Beni-Suef Univ. J. Basic Appl. Sci.
Agarwal, N., Bajpai, A., Srivastava, V., Gupta, S. P. 2013. A Quantitative Structure-
TRP988, TRP988 TRP931 TRP931, and one-one hydrophobic inter- Activity Relationship and Molecular Modeling Study on a Series of Biaryl
action with LEU958, ALA966, and VAL26. More also it was found Imidazole Derivatives Acting as H. Struct. Biol. 2013.
that the result obtained where better than the one proposed by Anand, K., Wakode, S., 2017. Development of drugs based on Benzimidazole
Heterocycle: Recent Advancement and Insights. IJCS 5 (2), 350–362.
other researchers. Chaudhary, S., Kapila, A., Sharma, R.B., Vashist, H., 2015. A review on peptic ulcer.
In addition, all the quinazolinone derivatives docked with H+/ Int. J. Recent Adv. Sci. Technol. 2, 6–10.
+
K -ATPase enzyme were better than even the standard anti-ulcer Drini, M., 2017. Peptic ulcer disease and non-steroidal anti-inflammatory drugs.
Austr. Prescriber 40, 91.
drug (omeprazole). The physicochemical parameter used for
Jalali-Heravi, M., Kyani, A., 2004. Use of computer-assisted methods for the
molecular docking and quantitative structure-affinity relationship modeling of the retention time of a variety of volatile organic compounds: a
(QSAR) in this study were essential parameters to look into espe- PCA-MLR-ANN approach. J. Chem. Inf. Comput. Sci. 44, 1328–1335.
cially in improving the chance of new anti-ulcer drug as quinazoli- Kennard, R.W., Stone, L.A., 1969. Computer aided design of experiments.
Technometrics 11, 137–148.
none analogues, our QSAR result with high correlation coefficient Konturek, P.C., Konturek, S.J., 2014. Peptic ulcer disease. Metabolism of Human
R2 of 0.8946 and molecular docking result of
9.3 kcal/mol Diseases. Springer.
666 M.baba Muh’d et al. / Journal of King Saud University – Science 32 (2020) 657–666

Li, Y.-R., Li, C., Liu, J.-C., Guo, M., Zhang, T.-Y., Sun, L.-P., Piao, H.-R., 2015. Synthesis Shamsuddeen, U., Yusha’u, M., Adamu, I., 2009. Helicobacter pylori: the causative
and biological evaluation of 1, 3-diaryl pyrazole derivatives as potential agent of peptic ulcer. Bayero J. Pure Appl. Sci. 2 (2), 79–83.
antibacterial and anti-inflammatory agents. Bioorg. Med. Chem. Lett. 25 (22), Tropsha, A., Gramatica, P., Gombar, V.K., 2003. The importance of being earnest:
5052–5057. validation is the absolute essential for successful application and interpretation
Noor, A., Qazi, N.G., Nadeem, H., Khan, A.-U., Paracha, R.Z., Ali, F., Saeed, A., 2017. of QSPR models. Mol. Inf. 22, 69–77.
Synthesis, characterization, anti-ulcer action and molecular docking evaluation Trott, O., Olson, A.J., 2010. AutoDock Vina: improving the speed and accuracy of
of novel benzimidazole-pyrazole hybrids. Chem. Cent. J. 11, 85. docking with a new scoring function, efficient optimization, and
Rajesh, R., Manikandan, A., Sivakumar, A., Ramasubbu, C., Nagaraju, N., 2017. multithreading. J. Comput. Chem. 31, 455–461.
Substituted methoxybenzyl-sulfonyl-1H-benzo [d] imidazoles evaluated as Utzeri, E., Usai, P., 2017. Role of non-steroidal anti-inflammatory drugs on intestinal
effective H+/K+-ATPase inhibitors and anti-ulcer therapeutics. Eur. J. Med. permeability and nonalcoholic fatty liver disease. World J. Gastroenterol. 23,
Chem. 139, 454–460. 3954.
Rakesh, K., Shantharam, C., Manukumar, H., 2016. Synthesis and SAR studies of Yap, C.W., 2011. PaDEL-descriptor: An open source software to calculate molecular
potent H+/K+-ATPase inhibitors of quinazolinone-schiff’s base analogues. descriptors and fingerprints. J. Comput. Chem. 32, 1466–1474.
Bioorg. Chem. 68, 1–8.